您的位置：山东大学 -> 科技期刊社 -> 《山东大学学报（理学版）》

《山东大学学报(理学版)》 ›› 2024, Vol. 59 ›› Issue (3): 14-26.doi: 10.6040/j.issn.1671-9352.7.2023.9950

•   • 上一篇    下一篇

模糊概念集的启发式构造方法及其推荐应用

刘忠慧1(),姜帅1,闵帆1,2,*()   

  1. 1. 西南石油大学计算机科学学院，四川 成都 610500
    2. 西南石油大学人工智能研究院，四川 成都 610500
  • 收稿日期:2023-04-28 出版日期:2024-03-20 发布日期:2024-03-06
  • 通讯作者: 闵帆 E-mail:lz_hui@126.com;minfan@swpu.edu.cn
  • 作者简介:刘忠慧(1980—)，女，教授，硕士生导师，硕士，研究方向为机器学习、形式概念分析与粗糙集等. E-mail: lz_hui@126.com
  • 基金资助:
    国家自然科学基金资助项目(61976245);中央引导地方科技发展专项资助项目(2021ZYD0003)

Heuristic construction method of fuzzy concept set and its recommended application

Zhonghui LIU1(),Shuai JIANG1,Fan MIN1,2,*()   

  1. 1. School of Computer Science, Southwest Petroleum University, Chengdu 610500, Sichuan, China
    2. Institute for Artificial Intelligence, Southwest Petroleum University, Chengdu 610500, Sichuan, China
  • Received:2023-04-28 Online:2024-03-20 Published:2024-03-06
  • Contact: Fan MIN E-mail:lz_hui@126.com;minfan@swpu.edu.cn

RichHTML

3

PDF (PC)

414

摘要：

针对模糊形式概念分析在推荐应用中难以用于大规模数据集的问题，提出了一种基于模糊概念集启发式构造的推荐方法。根据用户之间的相似度，为每个用户构建子背景，在子背景上采用新的启发式信息，分别以用户和项目为线索生成模糊概念。利用模糊概念内部信息，设计了融入用户权重的推荐置信度，实现了对用户的个性化推荐。在6个真实数据集上进行试验，本方法的推荐效率较高，与经典的协同过滤算法相比，在稀疏的数据集上能够取得更好的推荐效果。

关键词: 形式概念分析, 模糊概念, 概念构造, 推荐系统, 用户相似度

Abstract:

Aiming at the problem that fuzzy formal concept analysis is difficult to apply to large-scale datasets in recommendation applications, a recommendation method based on a heuristic construction of fuzzy concept set is proposed. Sub-contexts are constructed for each user based on the similarity between users. Then, new heuristic information is used on the sub-contexts to generate fuzzy concepts with users and items as clues, respectively. Finally, using the internal information of fuzzy concepts, a recommendation confidence integrated with user weights is designed to achieve personalized recommendations for users. The experimental results on six real datasets show that the proposed method has higher recommendation efficiency, and can achieve better recommendation results on sparse data sets compared with classical collaborative filtering algorithms.

Key words: formal concept analysis, fuzzy concept, concept construction, recommender system, user similarity

中图分类号: 

  • TP181

表1

主要符号"

符号 含义 符号 含义
U 用户集 C 概念集合
M 属性集 Hu=(Uu, Mu, Ĩu) 用户u子背景
I 二元关系 D(u1, u2) 用户相似度
K=(U, M, I) 形式背景 S(c) 模糊概念大小
Ĩ 模糊关系 wc(u1, u2) 用户权重
H=(U, M, Ĩ) 模糊形式背景 Rc(u, m) 推荐置信度
μ(*) 隶属度 Au 用户u线索集合
φ(*) 模糊集 Bm 项目m线索集合
c=(φ(A), B) 模糊概念 L 推荐矩阵

表2

形式背景"

UM
m1 m2 m3 m4 m5
u1 1 1 1 0 0
u2 1 1 1 1 1
u3 1 0 1 0 1
u4 1 1 0 1 0
u5 0 0 1 0 0

表3

模糊形式背景"

UM
m1 m2 m3 m4 m5
u1 1.0 0.2 0.6 0.0 0.0
u2 0.8 0.4 0.8 0.6 0.8
u3 0.6 0.0 0.2 0.0 0.4
u4 0.6 0.2 0.0 0.6 0.0
u5 0.0 0.0 0.6 0.0 0.0

表4

u1的子背景"

Uu1Mu1
m1 m2 m3
u1 1.0 0.2 0.6
u2 0.8 0.4 0.8
u3 0.6 0.0 0.2
u4 0.6 0.2 0.0

表5

在u1子背景下以用户为线索生成概念过程"

Au φ(Au**) Au* S(c)
{u1} {(u1, 0.2), (u2, 0.4)} {m1, m2, m3} 0.6
{u1, u2} {(u1, 0.2), (u2, 0.4)} {m1, m2, m3} 0.6
{u1, u2, u3} {(u1, 0.6), (u2, 0.8), (u3, 0.2)} {m1, m3} 0.9
{u1, u2, u3, u4} {(u1, 1.0), (u2, 0.8), (u3, 0.6), (u4, 0.6)} {m1} 0.7

表6

在u1子背景下以项目为线索生成概念过程"

Bu φ(Bu*) Bu** S(c)
{m1} {(u1, 1.0), (u2, 0.8), (u3, 0.6), (u4, 0.6)} {m1} 0.7
{m1, m2} {(u1, 0.2), (u2, 0.4), (u4, 0.2)} {m1, m2} 0.9
{m1, m2, m3} {(u1, 0.2), (u2, 0.4)} {m1, m2, m3} 0.6

表7

试验数据集"

数据集 用户数 项目数 评分数 稠密度
movielens-100k 943 1 682 100 000 0.063 0
movielens-1m 6 040 3 952 1 000 000 0.041 9
eachmovie-2ku 2 000 1 628 38 571 0.011 8
eachmovie-3ku 3 000 1 628 59 247 0.012 1
filmtrust 1 508 2 071 35 497 0.011 3
jester-s 8 000 100 577 379 0.721 7

图1

不同θ对概念生成时间与F1的影响"

图2

不同α对F1的影响"

图3

不同λ对F1的影响"

表8

RFCS与其他算法推荐效果对比"

数据集 算法 Pprecision Precall F1
movielens-100k RFCS 0.248 7 0.243 9 0.246 2
CSBR 0.192 5 0.239 3 0.213 2
kNN 0.199 4 0.344 2 0.252 5
IBCF 0.203 8 0.406 1 0.271 4
GreConD-kNN 0.199 6 0.342 4 0.252 1
NCF 0.194 6 0.335 2 0.246 1
movielens-1m RFCS 0.216 0 0.186 4 0.200 0
CSBR 0.142 1 0.204 0 0.167 6
kNN 0.188 3 0.339 0 0.242 0
IBCF 0.200 1 0.394 2 0.265 5
GreConD-kNN 0.188 5 0.338 0 0.242 1
NCF 0.139 5 0.359 3 0.200 8
eachmovie-2ku RFCS 0.223 6 0.251 8 0.236 8
CSBR 0.187 5 0.227 2 0.205 2
kNN 0.201 4 0.272 9 0.231 8
IBCF 0.148 6 0.352 0 0.208 9
GreConD-kNN 0.199 0 0.269 8 0.229 1
NCF 0.129 7 0.317 8 0.184 1
eachmovie-3ku RFCS 0.231 5 0.271 3 0.249 7
CSBR 0.203 0 0.236 2 0.218 3
kNN 0.112 3 0.276 8 0.159 6
IBCF 0.151 0 0.389 4 0.217 7
GreConD-kNN 0.104 3 0.247 8 0.146 6
NCF 0.131 9 0.363 2 0.193 4
filmtrust RFCS 0.486 2 0.477 6 0.481 4
CSBR 0.360 2 0.641 7 0.461 2
kNN 0.438 6 0.441 8 0.441 0
IBCF 0.182 6 0.532 8 0.271 7
GreConD-kNN 0.467 9 0.406 6 0.435 0
NCF 0.207 9 0.742 2 0.324 7
jester-s RFCS 0.714 8 0.846 9 0.775 3
CSBR 0.572 7 0.839 7 0.681 0
kNN 0.880 3 0.902 9 0.891 5
IBCF 0.891 0 0.851 7 0.870 9
GreConD-kNN 0.895 2 0.614 0 0.728 4
NCF 0.971 0 0.853 4 0.908 4
1 WILLE R. Restructuring lattice theory: an approach based on hierarchies of concepts[C]//Formal Concept Analysis: 7th International Conference. Berlin: Springer, 2009: 314-339.
2 KUZNETSOV S O. Machine learning and formal concept analysis[C]//Concept Lattices: Second International Conference on Formal Concept Analysis. Berlin: Springer, 2004: 287-312.
3 KANG Xiangping , LI Deyu , WANG Suge . A multi-instance ensemble learning model based on concept lattice[J]. Knowledge-Based Systems, 2011, 24 (8): 1203- 1213.
doi: 10.1016/j.knosys.2011.05.010
4 KUMAR C A . Fuzzy clustering-based formal concept analysis for association rules mining[J]. Applied Artificial Intelligence, 2012, 26 (3): 274- 301.
doi: 10.1080/08839514.2012.648457
5 TALLAM S , GUPTA N . A concept analysis inspired greedy algorithm for test suite minimization[J]. ACM SIGSOFT Software Engineering Notes, 2005, 31 (1): 35- 42.
6 ZOU Caifeng , ZHANG Daqiang , WAN Jiafu , et al. Using concept lattice for personalized recommendation system design[J]. IEEE Systems Journal, 2015, 11 (1): 305- 314.
7 陈昊文, 王黎明, 张卓. 基于概念邻域的Top-N推荐算法[J]. 小型微型计算机系统, 2017, 38 (11): 2553- 2559.
doi: 10.3969/j.issn.1000-1220.2017.11.025
CHEN Haowen , WANG Liming , ZHANG Zhuo . Top-N recommendation algorithm based on conceptual neighborhood[J]. Journal of Chinese Computer Systems, 2017, 38 (11): 2553- 2559.
doi: 10.3969/j.issn.1000-1220.2017.11.025
8 ZADEH L A . Fuzzy sets[J]. Information and Control, 1965, 8 (3): 338- 353.
doi: 10.1016/S0019-9958(65)90241-X
9 QUAN T T, HUI S C, CAO T H. A fuzzy FCA-based approach for citation-based document retrieval[C]//IEEE Conference on Cybernetics and Intelligent Systems, 2004. Piscataway: IEEE, 2004, 1: 578-583.
10 BURUSCO J A , FUENTES-GONZÁLEZ R . The study of the L-fuzzy concept lattice[J]. Mathware and Soft Computing, 1994, 1 (3): 209- 218.
11 刘宗田, 强宇, 周文, 等. 一种模糊概念格模型及其渐进式构造算法[J]. 计算机学报, 2007, 30 (2): 184- 188.
doi: 10.3321/j.issn:0254-4164.2007.02.004
LIU Zongtian , QIANG Yu , ZHOU Wen , et al. Fuzzy concept lattice model and its incremental construction algorithm[J]. Chinese Journal of Computers, 2007, 30 (2): 184- 188.
doi: 10.3321/j.issn:0254-4164.2007.02.004
12 KRAJCI S . Cluster based efficient generation of fuzzy concepts[J]. Neural Network World, 2003, 13 (5): 521- 530.
13 YAHIA S B , AROUR K , SLIMANI A , et al. Discovery of compact rules in relational databases[J]. Information Science Journal, 2000, 4 (3): 497- 511.
14 ZHANG Wenxiu , MA Jianmin , FAN Shiqing . Variable threshold concept lattices[J]. Information Sciences, 2007, 177 (22): 4883- 4892.
doi: 10.1016/j.ins.2007.05.031
15 BOFFA S, DE MAIO C, DI NOLA A, et al. Unifying fuzzy concept lattice construction methods[C]//2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE). Piscataway: IEEE, 2016: 209-216.
16 SHAO Mingwen , LEUNG Y , WANG Xizhao , et al. Granular reducts of formal fuzzy contexts[J]. Knowledge-Based Systems, 2016, 114, 156- 166.
doi: 10.1016/j.knosys.2016.10.010
17 MI Yunlong , SHI Yong , LI Jinhai , et al. Fuzzy-based concept learning method: exploiting data with fuzzy conceptual clustering[J]. IEEE Transactions on Cybernetics, 2020, 52 (1): 582- 593.
18 XU Weihua , GUO Doudou , QIAN Yuhua , et al. Two-way concept-cognitive learning method: a fuzzy-based progressive learning[J]. IEEE Transactions on Fuzzy Systems, 2022, 31 (6): 1- 15.
19 SINGH P K , KUMAR C A , LI Jinhai . Knowledge representation using interval-valued fuzzy formal concept lattice[J]. Soft Computing, 2016, 20, 1485- 1502.
doi: 10.1007/s00500-015-1600-1
20 FANG Peici, ZHENG Siyao. A research on fuzzy formal concept analysis based collaborative filtering recommendation system[C]//2009 Second International Symposium on Knowledge Acquisition and Modeling. Piscataway: IEEE, 2009: 352-355.
21 张喜征, 蔡月月, 罗文. 基于模糊概念格的领先用户个性化知识推荐研究[J]. 科技管理研究, 2019, 39 (7): 183- 189.
ZHANG Xizheng , CAI Yueyue , LUO Wen . Research on personalized knowledge recommendation for leading users based on fuzzy concept lattice in innovation community[J]. Science and Technology Management Research, 2019, 39 (7): 183- 189.
22 LIU Zhonghui , ZHAO Qi , ZOU Lu , et al. A heuristic concept construction approach to collaborative recommendation[J]. International Journal of Approximate Reasoning, 2022, 146, 119- 132.
doi: 10.1016/j.ijar.2022.04.004
23 GANTER B , WILLE R . Formal concept analysis: mathematical foundations[M]. Berlin: Springer, 1999.
24 ZADEH L A . Fuzzy sets and information granularity[M]. Amsterdam: World Scientific Publisher, 1979.
25 李金海, 魏玲, 张卓, 等. 概念格理论与方法及其研究展望[J]. 模式识别与人工智能, 2020, 33 (7): 619- 642.
LI Jinhai , WEI Ling , ZHANG Zhuo , et al. Concept lattice theory and method and their research prospect[J]. Pattern Recognition and Artificial Intelligence, 2020, 33 (7): 619- 642.
26 POELMANS J , IGNATOV D I , KUZNETSOV S O , et al. Fuzzy and rough formal concept analysis: a survey[J]. International Journal of General Systems, 2014, 43 (2): 105- 134.
doi: 10.1080/03081079.2013.862377
27 于蒙, 何文涛, 周绪川, 等. 推荐系统综述[J]. 计算机应用, 2022, 42 (6): 1898- 1913.
YU Meng , HE Wentao , ZHOU Xuchuan , et al. Review of recommendation system[J]. Journal of Computer Applications, 2022, 42 (6): 1898- 1913.
28 BOUCHER-RYAN P D , BRIDGE D . Collaborative recommending using formal concept analysis[J]. Knowledge-Based Systems, 2006, 19 (5): 309- 315.
doi: 10.1016/j.knosys.2005.11.017
29 朵琳, 杨丙. 一种基于用户兴趣概念格的推荐评分预测方法[J]. 小型微型计算机系统, 2020, 41 (10): 2104- 2108.
DUO Lin , YANG Bing . Recommendation rating prediction based on user interest concept lattice[J]. Journal of Chinese Computer Systems, 2020, 41 (10): 2104- 2108.
30 MEZNI H , ABDELJAOUED T . A cloud services recommendation system based on fuzzy formal concept analysis[J]. Data & Knowledge Engineering, 2018, 116, 100- 123.
31 KUMAR C A , SINGH P K . Knowledge representation using formal concept analysis: a study on concept generation[M]. Hershey: IGI Global, 2014: 306- 336.
32 HARPER F M , KONSTAN J A . The movielens datasets: history and context[J]. ACM Transactions on Interactive Intelligent Systems, 2015, 5 (4): 1- 19.
33 PHUONG N D, PHUONG T M. Collaborative filtering by multi-task learning[C]//2008 IEEE International Conference on Research, Innovation and Vision for the Future in Computing and Communication Technologies. Piscataway: IEEE, 2008: 227-232.
34 GUO Guibing, ZHANG Jie, YORKE-SMITH N. A novel Bayesian similarity measure for recommender systems[C]//Proceedings of the Twenty-third International Joint Conference on Artificial Intelligence. Menlo Park: AAAI Press, 2013: 2619-2625.
35 GOLDBERG K , ROEDER T , GUPTA D , et al. Eigentaste: a constant time collaborative filtering algorithm[J]. Information Retrieval, 2001, 4, 133- 151.
36 IGNATOV D I, NENOVA E, KONSTANTINOVA N, et al. Boolean matrix factorisation for collaborative filtering: an FCA-based approach[C]//International Conference on Artificial Intelligence: Methodology, Systems, and Applications. Cham: Springer, 2014: 47-58.
37 HE Xiangnan, LIAO Lizi, ZHANG Hanwang, et al. Neural collaborative filtering[C]//Proceedings of the 26th International Conference on World Wide Web. Republic and Canton of Geneva: International World Wide Web Conferences Steering Committee, 2017: 173-182.
[1] 陈曜琦,徐伟华,蒋宗颖. 三支概念的恢复集[J]. 《山东大学学报(理学版)》, 2023, 58(12): 52-62.
[2] 韩培磊,魏玲,王振,赵思雨. FCA中的互补概念及其性质与生成[J]. 《山东大学学报(理学版)》, 2022, 57(8): 60-67.
[3] 唐洁,魏玲,任睿思,赵思雨. 基于可能属性分析的粒描述[J]. 《山东大学学报(理学版)》, 2021, 56(1): 75-82.
[4] 李金海,贺建君,吴伟志. 多粒度形式概念分析的类属性块优化[J]. 《山东大学学报(理学版)》, 2020, 55(5): 1-12.
[5] 刘营营,米据生,梁美社,李磊军. 三支区间集概念格[J]. 《山东大学学报(理学版)》, 2020, 55(3): 70-80.
[6] 李金海,吴伟志. 形式概念分析的粒计算方法及其研究展望[J]. 山东大学学报（理学版）, 2017, 52(7): 1-12.
[7] 黄淑芹,徐勇,王平水. 基于概率矩阵分解的用户相似度计算方法及推荐应用[J]. 山东大学学报（理学版）, 2017, 52(11): 37-43.
[8] 汤亚强, 范敏, 李金海. 三元形式概念分析下的认知系统模型及信息粒转化方法[J]. 山东大学学报（理学版）, 2014, 49(08): 102-106.
[9] 戚丽丽,孙静宇*,陈俊杰. 基于均模型的IBCF算法研究[J]. J4, 2013, 48(11): 105-110.
[10] 孙静宇,陈俊杰*,余雪丽,何秀. 基于CBR的协同Web搜索模型及应用研究[J]. J4, 2012, 47(5): 19-24.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
[1] 霍玉洪,季全宝. 一类生物细胞系统钙离子振荡行为的同步研究[J]. J4, 2010, 45(6): 105 -110 .
[2] 石长光 . Faddeev模型中的多孤立子解[J]. J4, 2007, 42(7): 38 -40 .
[3] 马继雄,江莉,祁驭矜,向凤宁,夏光敏 . 祁连龙胆愈伤组织和再生植株的生长及其两种药效成分分析[J]. J4, 2006, 41(6): 157 -160 .
[4] 李永明1, 丁立旺2. PA误差下半参数回归模型估计的r-阶矩相合[J]. J4, 2013, 48(1): 83 -88 .
[5] 董丽红1,2,郭双建1. Yetter-Drinfeld模范畴上的弱Hopf模基本定理[J]. J4, 2013, 48(2): 20 -22 .
[6] 程李晴1,2, 石巧连2. 一种新的混合共轭梯度算法[J]. J4, 2010, 45(6): 81 -85 .
[7] 王康 李华. 化学计量学方法用于蛤青注射色谱数据重叠峰的分辨[J]. J4, 2009, 44(11): 16 -20 .
[8] 陈 莉, . 非方广义系统带干扰抑制的奇异LQ次优控制问题[J]. J4, 2006, 41(2): 74 -77 .
[9] 程智1,2,孙翠芳2,王宁1,杜先能1. 关于Zn的拉回及其性质[J]. J4, 2013, 48(2): 15 -19 .
[10] 张爱平,李刚 . LR拟正规Ehresmann半群[J]. J4, 2006, 41(5): 44 -47 .
版权所有 © 2018 《山东大学学报(理学版)》编辑部　
地址: 山东省济南市山大南路27号山东大学中心校区（250100）　电话: +86-0531-88366917　E-mail: xblxb@sdu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发